Filtering Torchtext Dataset by Field

nlp
1

Hi,
Given the MNLI data set with something as as follows:

TEXT = data.Field(lower=True)
LABEL = data.LabelField(sequential=False)
GENRE = data.LabelField(sequential=False)
train, val, test = datasets.MNLI.splits(TEXT, LABEL, GENRE)

How do I filter the MNLI dataset to only include a particular genre? I only want to do training on a single genre, for example “Fiction”.